Digital PCR System for Absolute

Published:October 28,2011

ARTICLE https://www.wendangku.net/doc/6f9884025.html,/ac

High-Throughput Droplet Digital PCR System for Absolute Quantitation of DNA Copy Number

Benjamin J.Hindson,*,?Kevin D.Ness,?Donald A.Masquelier,?Phillip Belgrader,?Nicholas J.Heredia,?Anthony J.Makarewicz,?Isaac J.Bright,?Michael Y.Lucero,?Amy L.Hiddessen,?Tina C.Legler,?Tyler K.Kitano,?Michael R.Hodel,?Jonathan F.Petersen,?Paul W.Wyatt,?Erin R.Steenblock,?Pallavi H.Shah,?Luc J.Bousse,?Camille B.Troup,?Je ?rey C.Mellen,?Dean K.Wittmann,?

Nicholas G.Erndt,?Thomas H.Cauley,?Ryan T.Koehler,?Austin P.So,?Simant Dube,?Klint A.Rose,?Luz Montesclaros,?Shenglong Wang,?David P.Stumbo,?Shawn P.Hodges,?Steven Romine,?Fred https://www.wendangku.net/doc/6f9884025.html,anovich,?Helen E.White,?John F.Regan,?George A.Karlin-Neumann,?Christopher M.Hindson,?Serge Saxonov,?and Bill W.Colston ?

?

Bio-Rad Laboratories,Inc.,7068Koll Center Parkway,Pleasanton,California 94566,United States ?

National Genetics Reference Laboratory,Wessex Regional Genetics,Salisbury District Hospital,Odstock,Salisbury,Wiltshire,SP28BJ,United Kingdom

b

Supporting Information D

etection and quantitation of speci ?c nucleic acid sequences using PCR is fundamental to a large body of research and a growing number of molecular diagnostic tests.The ?rst genera-tion of PCR users performed end-point analysis by gel electro-phoresis to obtain qualitative results.The advent of real-time PCR spawned a second generation that enabled quantitation by monitoring the progression of ampli ?cation after each cycle using ?uorescence probes.In real-time PCR,quantitative infor-mation is obtained from the cycle threshold (C T ),a point on the analogue ?uorescence curve where the signal increases above background.External calibrators or normalization to endogen-ous controls are required to estimate the concentration of an unknown.Imperfect ampli ?cation e ?ciencies a ?ect C T values which in-turn limits the accuracy of this technique for absolute quantitation.

Early pioneers 1recognized that the combination of limiting dilution,end-point PCR,and Poisson statistics could yield an absolute measure of nucleic acid concentration,an approach that later became known as digital PCR.2In digital PCR,target DNA molecules are distributed across multiple replicate reactions at a level where there are some reactions that have no template and

others that have one or more template copies present.After ampli ?cation to the terminal plateau phase of PCR,reactions containing one or more templates yield positive end-points,whereas those without template remain negative.The number of target DNA molecules present can be calculated from the fraction of positive end-point reactions using Poisson statistics,according to eq 1,λ?àln e1àp T

e1T

where λis the average number of target DNA molecules per replicate reaction and p is the fraction of positive end-point reactions.From λ,together with the volume of each replicate PCR and the total number of replicates analyzed,an estimate of the absolute target DNA concentration is calculated.In digital PCR,the number of replicates,or partitions,largely de ?nes the dynamic range of target DNA quantitation,where an order of magnitude increase in the number of replicates yields

Received:August 3,2011Accepted:October 5,2011ABSTRACT:Digital PCR enables the absolute quantitation of nucleic acids in a sample.The lack of scalable and practical technologies for digital PCR implementation has hampered the widespread adoption of this inherently powerful technique.Here we describe a high-throughput droplet digital PCR (ddPCR)system that enables processing of ～2million PCR reactions using conventional TaqMan assays with a 96-well plate work ?ow.Three applications demonstrate that the massive partitioning a ?orded by our ddPCR system provides orders of magni-tude more precision and sensitivity than real-time PCR.First,we show the accurate measurement of germline copy number variation.Second,for rare alleles,we show sensitive detection of mutant DNA in a 100000-fold excess of wildtype background.Third,we demonstrate absolute quantitation of circulating fetal and maternal DNA from cell-free plasma.We anticipate this ddPCR system will allow researchers to explore complex genetic landscapes,discover and validate new disease associations,and de ?ne a new era of molecular

diagnostics.

approximately an order of magnitude increase in dynamic range. Increasing the number of partitions also improves precision and therefore enables resolution of small concentration di?erences between nucleic acid sequences in a sample.4This is analogous to the relationship between the number of pixels and the resolution of a digital image.As digital PCR relies on a binary end-point threshold to assign each replicate reaction as either positive or negative(one or zero,respectively),it can tolerate wide varia-tions in ampli?cation e?ciencies without a?ecting DNA copy number estimation.Despite its low-throughput and limited dynamic range,digital PCR by limiting dilution in microwell plates is still used today.3A practical and low-cost embodiment will unlock the potential of digital PCR and establish a third generation of PCR users and applications.

Currently there are two approaches used by commercially available digital PCR systems.The?rst approach uses micro-wells5or micro?uidic chambers6à8to split the sample into hundreds of nanoliter partitions.Micro?uidic chips simplify reaction setup but are challenging to scale to achieve high-throughput.The second approach,called BEAMing,9,10is based on emulsion PCR,where templates are clonally ampli?ed in the presence of beads.Post-PCR,the emulsion is broken to recover the beads,which are subsequently labeled with a?uorescent hybridization probe and read by conventional?ow-cytometry. BEAMing requires specialized heterogeneous assay schemes that add complexity to the work?ow thereby limiting its adoption to a few applications including rare allele detection and DNA methylation.11à13Overall,high costs,limited throughput,and complicated work?ows have hampered the adoption of digital PCR.

We have developed an approach that uses water-in-oil droplets14à16as the enabling technology to realize high-through-put digital PCR in a low-cost and practical format.Our approach takes advantage of simple micro?uidic circuits and surfactant chemistries to divide a20μL mixture of sample and reagents into ～20000monodisperse droplets(i.e.,partitions).These droplets support PCR ampli?cation of single template molecules using homogeneous assay chemistries and work?ows similar to those widely used for real-time PCR applications(i.e.,TaqMan).An automated droplet?ow-cytometer reads each set of droplets after PCR at a rate of32wells per hour.

’RESULTS AND DISCUSSION

The droplet digital PCR(ddPCR)work?ow requires the following steps(Figure1):Eight assembled PCR reactions,each comprising template,ddPCR Mastermix and TaqMan reagents, are loaded into individual wells of a single-use injection molded cartridge.Next,droplet generation oil containing stabilizing surfactants is loaded and the cartridge placed into the droplet generator.By application of vacuum to the outlet wells,sample and oil are drawn through a?ow-focusing junction where mono-disperse droplets are generated at a rate of～1000per second. The surfactant-stabilized droplets?ow to a collection well where they quickly concentrate due to density di?erences between the

oil and aqueous phases,forming a packed bed above the excess oil.The densely packed droplets are pipet transferred to a conventional96-well PCR plate and thermal cycled to end-point. After thermal cycling,the plate is transferred to a droplet reader. Here,droplets from each well are aspirated and streamed toward the detector where,en route,the injection of a spacer?uid separates and aligns them for single-?le simultaneous two-color detection.TaqMan assays provide speci?c duplexed detection of target and reference genes.All droplets are gated based on detector peak width to exclude rare outliers(e.g.,doublets, triplets).Each droplet has an intrinsic?uorescence signal result-ing from the imperfect quenching of the?uorogenic probes enabling detection of negative droplets.

For droplets that contain Figure1.Droplet digital PCR work?ow:(a)Samples and droplet generation oil are loaded into an eight-channel droplet generator cartridge.(b)A vacuum is applied to the droplet well,which draws sample and oil through a?ow-focusing nozzle where monodisperse1nL droplets are formed.In<2min,eight samples are converted into eight sets of20000droplets.(c)The surfactant-stabilized droplets are pipet transferred to a96-well PCR plate.(d)Droplet PCR ampli?cation to end-point(35à45cycles)is performed in a conventional thermal cycler.

(e)The plate is loaded onto a reader which sips droplets from each well and streams them single-?le past a two-color detector at the rate of ～1000per second.(f)Droplets are assigned as positive or negative based on their?uorescence amplitude.The number of positive and negative droplets in each channel is used to calculate the concentration of the target and reference DNA sequences(see eq1)and their Poisson-based95%con?dence intervals.

template,speci ?c cleavage of TaqMan probes generates a strong ?uorescence signal.On the basis of ?uorescence amplitude,a simple threshold assigns each droplet as positive or negative.As the droplet volume is known,the fraction of positive droplets is then used to calculate the absolute concentration of the target sequence.For 20000droplets,the dynamic range for absolute quantitation spans from a single copy up to ～100000copies.For human genomic DNA,this equates to an input DNA mass ranging from 3.3fg to 330ng per 20μL reaction.As templates are randomly distributed across the droplet partitions,a Poisson correction extends the dynamic range into the realm where on average there are multiple copies per droplet.Statistical models are applied to calculate con ?dence limits of the concentration estimates and their ratios.4,17

To demonstrate the immediate utility of this ddPCR system,we present data on three application areas of increasing interest to researchers:determination of copy number variation (CNV),detection of rare alleles and the absolute quantitation of circulat-ing DNA in cell-free plasma.Each application was selected to highlight a distinct advantage that massive droplet partitioning a ?ords to digital PCR.For CNV,the large number of replicates provides su ?cient precision to accurately measure copy number states.For the detection of rare alleles,partitioning the target mutant DNA away from highly homologous wildtype DNA increases sensitivity.Finally,droplet partitioning enables accurate quantitation of nucleic acids from clinical samples over a wide dynamic range without external calibrators or endogenous https://www.wendangku.net/doc/6f9884025.html,Vs are deletions and ampli ?cations of genome segments ranging from hundreds to millions of base pairs in length that have been implicated in a broad spectrum of human disease.18Microarrays and the next-generation sequencing technologies have enabled and accelerated the discovery of new CNVs,19thereby further increasing the need for a high-throughput,low-cost approach to making precise CNV measurements with increased dynamic range for validation and follow-up studies.Although microarray technologies are valuable tools for CNV discovery,20they have limited dynamic range and are expensive to scale to large numbers of samples for population studies.Multiplex ligation-dependent probe ampli ?cation (MLPA)21is an assay that allows resolution of deletions or duplications for up to 40targets but requires selection from a prede ?ned test menu or extensive upfront assay optimization for new target https://www.wendangku.net/doc/6f9884025.html,V investigators using methods based on real-time PCR have reported technical di ?culty obtaining accurate copy number measurements.22Real-time PCR measurements are inherently imprecise,and copy number estimates can drift between cases and controls.

We measured the germline copy number variation of HapMap samples by ddPCR.Because increases in gene copy number are often the result of tandem gene duplications,we used restriction enzymes to predictably and e ?ciently separate linked copies of the target gene such that each sequence is encapsulated into its own droplet and counted separately.Restriction enzymes were selected to cut either side of the amplicon sequences avoiding known mutation sites 23and methylation sensitivities.Physically shearing DNA using ultrasound or micro ?uidic devices is less attractive as it reduces the amount of target that can be ampli ?ed by PCR and can be challenging to implement in high-throughput without specialized equipment.Preampli ?cation,an alternative strategy for separation of linked copies 24has the potential to introduce bias between the target and reference genes.

Seven HapMap samples were screened for CNVs for three target genes.Each ddPCR reaction contained duplex TaqMan assay reagents for the target and reference genes.For MRGPRX1,the copy number states from 1up to 6were completely resolved from the results of a single well for each sample (Figure 2a).Lower CNV states for CYP2D6and Chromosome X were also easily resolved,as shown.For 13HapMap samples,our system estimated the copy number of CCL3L1,a gene associated with HIV-1/AIDS susceptibility 18(Figure 2b).For DNA sample NA18507,next-generation sequencing estimated the CCL3L1copy number to be 5.725whereas our ddPCR system estimated 6.05.The estimate of 5.7is likely due to under-sampling since the billions of reads of a next-generation sequencing run are dis-tributed across the entire genome giving an average read-depth of only 30?.Thus,once target genes have been identi ?ed,greater precision can readily be achieved with ddPCR since the number of reads can be scaled almost arbitrarily.The current ddPCR system can achieve read depths of up to 20000?for two genes from a single well.These data show that our ddPCR system is well suited for CNV population studies as it enables large numbers of samples to be tested against smaller

sets of genes.

Figure 2.Determination of copy number variation states by droplet digital PCR.(a)Measured copy number variation states in HapMap samples for MRGPRX1,Chromosome X ,CYP2D6,and (b)CCL3L1.(c)Correlation of measured copy number alterations of GRB7and ERBB2in DNA extracted from normal and tumorous breast tissues.Each marker represents a CNV measurement from a single ddPCR well of ～20000droplets.Error bars indicate the Poisson 95%con ?dence intervals for each copy number determination.

Sample heterogeneity can attenuate the measurement of copy number ampli ?cations,which requires more precise measure-ments to discriminate smaller di ?erences from normal.Somatic copy number alteration is the hallmark of many cancers.Without high-throughput technology for precise copy number quantita-tion,pathologists use ?uorescent in situ hybridization (FISH)for diagnosing ampli ?cations and deletions as this technique a ?ords single-cell resolution.FISH and related techniques are expensive,laborious,and subject to large losses in performance due to other analytical factors.26Speci ?c ampli ?cations de ?ne tumor subtypes and guide therapy.For example,Her2positive breast tumors respond to Trastuzumab (Herceptin).For a set of normal and tumor breast tissue samples,the measured copy numbers of ERBB2and GRB7correlated with the exception of two samples that showed lower GRB7ampli ?cation (Figure 2c).These results were expected as the GRB7gene is part of the HER2amplicon

and is coampli ?ed in almost all breast tumors with 17q11-21ampli ?cation.27,28This ddPCR method provides the ability to scale the number of partitions by combining replicate wells to resolve ?ne copy number di ?erences in heterogeneous mixtures and could foreseeably form the basis of more e

?cient diagnostic tests.

Figure 3.Detection of the BRAF V600E rare mutant allele in the presence of homologous wildtype DNA by droplet digital PCR.Serial dilutions of the mutant cell line DNA were prepared in a constant background of wildtype human genomic DNA.Droplet partitioning reduces competitive ampli ?cation e ?ects allowing detection down to 0.001%mutant fraction,1000times lower than real-time PCR.The mutant cell line contains 35%BRAF V600E,as

measured by ddPCR.

Figure 4.Absolute quantitation of circulating fetal and maternal DNA from cell-free plasma for male and female fetuses.(a)Quantitation of fetal DNA concentration using SRY (red bar)and hypermethylated RASSF1(blue bar).The RASSF1gene of circulating fetal DNA is hypermethylated whereas maternal DNA is hypomethylated.Methyla-tion sensitive restriction enzymes selectively digested away the hypo-methylated fraction,leaving the hypermethylated fetal DNA that was quanti ?ed.(b)Quantitation of total DNA concentration (black bar)represented as the weighted average from six independent assay measurements including undigested RASSF1and β-actin as well as RNaseP and TERT .(c)Fetal loads as determined from the ratio of SRY to total (male fetuses only)and RASSF1to total (male and female fetuses).For male fetuses,the Pearson ’s correlation coe ?cient between SRY and RASSF1fetal loads was 97.3%.Fetal DNA is not completely hypermethylated;therefore,the RASSF1fetal loads measured for some samples are lower than those determined using SRY .Error bars represent the Poisson 95%con ?dence intervals of the concentration or the ratio in the case of fetal load estimates.

The second application demonstrates improved detection of rare mutant alleles by drastically reducing competitive PCR pro-cesses that occur in the presence of a highly homologous wild-type DNA background.With careful optimization,real-time PCR assays can detect down to the1%mutant fraction.With the same assays,ddPCR partitions the competing background away from the mutant,e?ectively increasing the average mutant-to-wild-type ratio by20000times.On average,the e?ective enrichment of the mutant molecules per PCR reaction is proportional to the number of sample partitions used.For a duplex TaqMan assay targeting the BRAF V600E mutation,29we show droplet parti-tioning detects0.001%mutant fraction,1000times lower than real-time PCR(Figure3and Supplementary Table1and Supple-mentary Figure1in the Supporting Information).With depen-dence on the amount recovered from clinical samples,more DNA can be loaded into the ddPCR system to push the detection limits down to even lower levels.This approach enables re-searchers to measure extremely low levels of mutant that could in turn lead to the improved detection of minimal residual disease and less invasive diagnostics.

We next evaluated the ability of this ddPCR system to quan-titate DNA in clinical samples.Circulating DNA in cell-free plasma30has received increasing levels of attention as a sample type for developing noninvasive prenatal31and oncology32 diagnostics.The cell-free DNA in plasma is highly fragmented33 and present at low levels,which present challenges for quantita-tion.We enumerated fetal and total DNA in maternal cell-free plasma.For19maternal plasma samples taken between10and 20weeks gestational age,the level of fetal(Figure4a)and total DNA(Figure4b)were measured for both male and female fetuses.A selective methylation-sensitive digest enabled the low-levels of hypermethylated RASSF1fetal DNA34to be accurately quanti?ed using our ddPCR system.With an absolute measure of SRY,RASSF1,and total DNA concentrations,the fetal load for each sample was calculated(Figure4c).For male fetuses,a correlation of93.7%between the hypermethylated RASSF1fetal DNA and SRY fetal loads provided con?dence in the estimates for female fetuses.On the basis of RASSF1alone,fetal loads ranged from2.1to11.9%and were in general agreement with those data collected by next-generation sequencing35that is currently limited to estimating fetal loads from male fetuses. This application demonstrates the capability of absolute quanti-tation of highly fragmented cell-free DNA in clinical samples.

Overall,these data show that ddPCR o?ers a practical solution to realize precise estimates of DNA copy number with high-throughput.We anticipate this system will unlock the inherent power of digital PCR to more researchers for many applications.

’EXPERIMENTAL SECTION

Droplet Digital PCR Workflow.The ddPCR workflow was described in Figure1.The TaqMan PCR reaction mixture was assembled from a2?ddPCR Mastermix(Bio-Rad),20?primer, and probes(final concentrations of900and250nM,respectively) and template(variable volume)in a final volume of20μL.Each assembled ddPCR reaction mixture was then loaded into the sample well of an eight-channel disposable droplet generator cartridge(Bio-Rad).A volume of60μL of droplet generation oil (Bio-Rad)was loaded into the oil well for each channel.The cartridge was placed into the droplet generator(Bio-Rad).The cartridge was removed from the droplet generator,where the droplets that collected in the droplet well were then manually transferred with a multichannel pipet to a96-well PCR plate.The plate was heat-sealed with a foil seal and then placed on a conventional thermal cycler and amplified to the end-point (40à55cycles).After PCR,the96-well PCR plate was loaded on the droplet reader(Bio-Rad),which automatically reads the droplets from each well of the plate(32wells/h).Analysis of the ddPCR data was performed with QuantaSoft analysis software(Bio-Rad)that accompanied the droplet reader. Determination of Copy Number Variation in HapMap Samples.For MRGPRX1,ChromosomeX,and CYP2D6,4.4μg of each purified human genomic DNA sample(Coriell)was digested with10units of RsaI(NEB)in50μL for1h at37°C. The digest was diluted8-fold to400μL with TE buffer(pH8.0) then33ng(3μL)was assayed per20μL ddPCR reaction.For CCL3L1,815ng of each purified human genomic DNA sample (Coriell)was digested with7.5units of MseI(NEB)in10μL for 1h at37°C.The digest was diluted3.5-fold to35μL with TE buffer and then69ng(3μL)was assayed per20μL ddPCR reaction.MRGPRX1assay sequences36were(forward primer) 50-TTAAGCTTCATCAGTATCCCCCA-30,(reverse primer) 50-CAAAGTAGGAAAACATCATCACAGGA-30,and(probe) 6FAM-ACCATCTCTAAAATCCT-MGBNFQ.Chromosome X assay sequences37were(forward primer)50-GATGAGGAAGG-CAATGATCC-30,(reverse primer)50-TTGGCTTTTACCA-AATAGGG-30,and(probe)50-FAM-TGTTTCTCTCTGCC-TGCACTGG-BHQ1-30(Integrated DNA Technologies).The CYPD2D6(Hs00010001_cn)was purchased as a20?premix of primers and FAM-MGBNFQ probe(Applied Biosystems).Mod-ified CCL3L1assay sequences19were(forward primer)50-GGG-TCCAGAAATACGTCAGT-30,(reverse primer)50-CATGTT-CCCAAGGCTCAG-30,and(probe)6FAM-TTCGAGGCC-CAGCGACCTCA-MGBNFQ.All CNV assays were duplexed with an RPP30reference assay(forward primer)50-GATTTG-GACCTGCGAGCG-30,(reverse primer)50-GCGGCTGTCT-CCACAAGT-30,and(probe)VIC-CTGACCTGAAGGCTCT-MGBNFQ.Thermal cycling conditions were95°C?10min (1cycle),94°C?30s and60°C?60s(40cycles),98°C?10min(1cycle),and12°C hold.

Determination of GRB7and ERBB2Copy Number Altera-tions.Purified DNA(20ng)from each normal and tumorous breast tissue sample(D8235086-1,Biochain)was digested with 0.2units of NlaIII in10μL for1h at37°C.The restricted DNA was added directly to ddPCR Mastermix at8.8ng(4.4μL)per 20μL of ddPCR reaction.ERBB2(Hs02803918_cn)and GRB7 (Hs02139994_cn)assays were purchased as20?premixes of primers and FAM-MGBNFQ probe(Applied Biosystems)and duplexed with the RPP30reference assay described above. Thermal cycling conditions were95°C?10min(1cycle), 94°C?30s and60°C?60s(40cycles),98°C?10min (1cycle),and12°C hold.

Rare Allele Detection.A dilution series of BRAF V600E mutant DNA(HTB-38D)from a HT-29cell line(ATCC)was prepared in a high,constant background(5000copies/μL)of wildtype DNA(NA19205,Coriell).For ddPCR,when the concentration of intact human genomic DNA is>66ng/20μL reaction,the accompanying increase in viscosity can cause the average droplet volume to change,which in turn could affect the accuracy of DNA quantitation.Therefore,for samples of this nature,restriction enzyme digestion is recommended to frag-ment the DNA and reduce solution viscosity.In our experience, once fragmented,the human genomic DNA concentration can

exceed1μg/20μL reaction without affecting the average droplet volume.Therefore,prior to ddPCR,each sample of the dilution series was digested with40U of Hae III(NEB)in100μL con-taining1?NEB buffer4and BSA.The BRAF V600E/wildtype duplex TaqMan assay used common primers(forward)50-CTA-CTGTTTTCCTTTACTTACTACACCTCAGA-30,(reverse) 50-ATCCAGACAACTGTTCAAACTGATG-30,and specific probes(BRAF V600E)6FAM-TAGCTACAGAGAAATC-MG-BNFQ and(wildtype)VIC-CTAGCTACAGTGAAATC-MGB-NFQ.Eight ddPCR wells were used for each sample of the dilution series.Thermal cycling conditions were95°C?10min (1cycle),94°C?30s and62.7°C?60s(55cycles),and 12°C hold.

Quantitation of Cell-Free Fetal and Total DNA in Maternal Plasma.Whole blood(3?10mL)was collected(ProMedDx) from healthy pregnant donors,between10and20weeks of gestational age,by venipuncture into cell-free DNA BCT tubes (Streck)according to the manufacturer’s instructions.Fetus gender was determined by ultrasound within6weeks of sample collection.The tubes were stored for up to48h at room temperature then shipped overnight at4°C to Bio-Rad where they were processed upon receipt.The whole blood was cen-trifuged for10min at1600g,the supernatant removed and transferred to a new tube,centrifuged for10min at16000g,the supernatant removed,and transferred to a new tube,then the cell-free plasma was stored atà80°C.Cell-free plasma(5mL) was thawed and cell-free DNA isolated using the QIAmp Circulating Nucleic Acid Kit(Qiagen)according to the manu-factuer's protocol and eluted in AVE buffer(150μL).A portion of the eluate(99μL)was subjected to a single-tube digest containing Hha I(30U),Hpa I I(60U),and BstU I(30U)in1?NEB buffer4in a total volume of120μL.A second portion of the eluate(33μL)was used in a no-digest control mixture where restriction enzymes were substituted for water.The mixtures were incubated for37°C for2h,60°C for2h,then65°C for 20min.The restriction enzyme digested mixture was split and subjected to three ddPCR duplexed assays of SRY/TERT, RASSF1/RNaseP,and RASSF1/β-actin.The restriction enzyme mixture cuts unmethylated RASSF1andβ-actin TaqMan tem-plates but not SRY,RNaseP,or TERT.The no-digest control mixture was split and subjected to two ddPCR duplexed assays of RASSF1/RNaseP and RASSF1/β-actin.β-Actin is hypomethy-lated in both fetal and maternal DNA and is completely digested by the enzyme cocktail.

RASSF134and SRY37assays were reported previously.RNaseP and TERT copy number reference assays were purchased com-mercially(Applied Biosystems).Theβ-actin assay was modi?ed from Chan et al.(forward primer)50-GCAAAGGCGAGGC-TCTGT-30,(reverse primer)50-CGTTCCGAAAGTTGCCTT-TTATGG-30,and(probe)VIC-ACCGCCGAGACCGCGTC-MGBNFQ.For RASSF1/RNaseP and RASSF1/β-actin duplexes, 1?GC-Rich Solution(Roche)was used as a component of the assembled ddPCR reaction mixtures.Thermal cycling conditions were95°C?10min(1cycle),95°C?30s and60°C?60s (45cycles),and4°C hold.

For each sample,six independent assay measurements of total DNA concentration(G.E/mL)were made from one TERT,one β-actin,two RASSF1,and two RNaseP assays.Each assay mea-surement comprised data from seven replicate ddPCR wells.We combined the droplet counts(positive and negative)from all seven replicate wells to yield a single“metawell”.The concentra-tion and con?dence intervals for each of the6measurement metawells were computed.4The appropriate dilution factors were applied to yield total cell-free DNA concentration(G.E./mL) and the con?dence interval is scaled accordingly.The weighted mean of the six total measurements was calculated,where weights are inverses of con?dence interval variances of these measurements.For digested RASSF1,there are two independent assay measurements,which are also combined in the same manner.For SRY,there is one measurement that was used directly,with scaling by a factor of2to account for haploidy. Fetal load is then computed as a ratio with the associated Poisson 95%con?dence intervals.

’ASSOCIATED CONTENT

b Supporting Information.Additional information as noted in text.This material is available free of charge via the Internet at https://www.wendangku.net/doc/6f9884025.html,.

’AUTHOR INFORMATION

Corresponding Author

*E-mail:ben_hindson@https://www.wendangku.net/doc/6f9884025.html,.

’ACKNOWLEDGMENT

The project described was supported by Grant Number R01EB010106from the National Institute of Biomedical Ima-ging and Bioengineering.The content is solely the responsibility of the authors and does not necessarily represent the o?cial views of the National Institute of Biomedical Imaging and Bioengineering or the National Institutes of Health.

’REFERENCES

(1)Sykes,P.J.;Neoh,S.H.;Brisco,M.J.;Hughes,E.;Condon,J.; Morley,A.A.Biotechniques1992,13,444–449.

(2)Vogelstein,B.;Kinzler,K.W.Proc.Natl.Acad.Sci.U.S.A.1999, 96,9236–9241.

(3)McCaughan,F.;Dear,P.H.J.Pathol.2010,220,297–306.

(4)Dube,S.;Qin,J.;Ramakrishnan,R.PLoS One2008,3,e2876.

(5)Morrison,T.;Hurley,J.;Garcia,J.;Yoder,K.;Katz,A.;Roberts,

D.;Cho,J.;Kanigan,T.;Ilyin,S.

E.;Horowitz,D.;Dixon,J.M.;Brenan,

C.J.Nucleic Acids Res.2006,34,e123.

(6)Warren,L.;Bryder,D.;Weissman,I.L.;Quake,S.R.Proc.Natl. Acad.Sci.U.S.A.2006,103,17807–17812.

(7)Ottesen,E.A.;Hong,J.W.;Quake,S.R.;Leadbetter,J.R.Science 2006,314,1464–1467.

(8)Fan,H.C.;Quake,S.R.Anal.Chem.2007,79,7576–7579.

(9)Dressman,D.;Yan,H.;Traverso,G.;Kinzler,K.W.;Vogelstein,

B.Proc.Natl.Acad.Sci.U.S.A.2003,100,8817–8822.

(10)Diehl,F.;Li,M.;He,Y.;Kinzler,K.W.;Vogelstein,B.; Dressman,D.Nat.Methods2006,3,551–559.

(11)Diehl,F.;Diaz,L.A.,Jr.Curr.Opin.Oncol.2007,19,36–42.

(12)Li,M.;Diehl,F.;Dressman,D.;Vogelstein,B.;Kinzler,K.W. Nat.Methods2006,3,95–97.

(13)Li,M.;Chen,W.D.;Papadopoulos,N.;Goodman,S.N.; Bjerregaard,N.C.;Laurberg,S.;Levin,B.;Juhl,H.;Arber,N.;Moinova, H.;Durkee,K.;Schmidt,K.;He,Y.;Diehl,F.;Velculescu,V.E.;Zhou,S.; Diaz,L.A.,Jr.;Kinzler,K.W.;Markowitz,S.D.;Vogelstein,B.;Nat. Biotechnol.2009,27(9),858à863.

(14)Beer,N.R.;Hindson,B.J.;Wheeler,E.K.;Hall,S.B.;Rose, K.A.;Kennedy,I.M.;Colston,B.W.Anal.Chem.2007,79,8471–8475.

(15)Beer,N.R.;Wheeler,E.K.;Lee-Houghton,L.;Watkins,N.; Nasarabadi,S.;Hebert,N.;Leung,P.;Arnold,D.W.;Bailey,C.G.; Colston,B.W.Anal.Chem.2008,80,1854–1858.

(16)Kiss,M.M.;Ortoleva-Donnelly,L.;Beer,N.R.;Warner,J.; Bailey,C.G.;Colston,B.W.;Rothberg,J.M.;Link,D.R.;Leamon,J.H. Anal.Chem.2008,80,8975–8981.

(17)Weaver,S.;Dube,S.;Mir,A.;Qin,J.;Sun,G.;Ramakrishnan, R.;Jones,R.C.;Livak,K.J.Methods2010,50,271–276.

(18)Gonzalez,E.;Kulkarni,H.;Bolivar,H.;Mangano,A.;Sanchez, R.;Catano,G.;Nibbs,R.J.;Freedman,B.I.;Quinones,M.P.;Bamshad, M.J.;Murthy,K.K.;Rovin,B.H.;Bradley,W.;Clark,R.A.;Anderson, S.A.;O’Connell R,J.;Agan,B.K.;Ahuja,S.S.;Bologna,R.;Sen,L.; Dolan,M.J.;Ahuja,S.K.Science2005,307,1434–1440.

(19)Sudmant,P.H.;Kitzman,J.O.;Antonacci,F.;Alkan,C.;Malig, M.;Tsalenko,A.;Sampas,N.;Bruhn,L.;Shendure,J.;Eichler,E.E. Science2010,330,641–646.

(20)Carter,N.P.Nat.Genet.2007,39,S16–21.

(21)Schouten,J.P.;McElgunn,C.J.;Waaijer,R.;Zwijnenburg,D.; Diepvens,F.;Pals,G.Nucleic Acids Res.2002,30,e57.

(22)Aldhous,M.C.;Abu Bakar,S.;Prescott,N.J.;Palla,R.;Soo,K.; Mans?eld,J.C.;Mathew,C.G.;Satsangi,J.;Armour,J.A.Hum.Mol. Genet.2010,19,4930–4938.

(23)Sherry,S.T.;Ward,M.H.;Kholodov,M.;Baker,J.;Phan,L.; Smigielski,E.M.;Sirotkin,K.Nucleic Acids Res.2001,29,308–311. (24)Qin,J.;Jones,R.C.;Ramakrishnan,R.Nucleic Acids Res.2008, 36,e116.

(25)Alkan,C.;Kidd,J.M.;Marques-Bonet,T.;Aksay,G.;Anto-nacci,F.;Hormozdiari,F.;Kitzman,J.O.;Baker,C.;Malig,M.;Mutlu, O.;Sahinalp,S.C.;Gibbs,R.A.;Eichler,E.E.Nat.Genet.2009, 41,1061–1067.

(26)Bartlett,J.M.;Ibrahim,M.;Jasani,B.;Morgan,J.M.;Ellis,I.; Kay,E.;Connolly,Y.;Campbell,F.;O’Grady,A.;Barnett,S.;Miller,K. Amer.J.Clin.Pathol.2009,131,106–111.

(27)Kauraniemi,P.;Kuukasjarvi,T.;Sauter,G.;Kallioniemi,A.Am. J.Pathol.2003,163,1979–1984.

(28)Luoh,S.W.Cancer Genet.Cytogenet.2002,136,43–47.

(29)Benlloch,S.;Paya,A.;Alenda,C.;Bessa,X.;Andreu,M.;Jover, R.;Castells,A.;Llor,X.;Aranda,F.I.;Massuti,B.J.Mol.Diagn.2006, 8,540–543.

(30)Lo,Y.M.;Corbetta,N.;Chamberlain,P.F.;Rai,V.;Sargent,

I.L.;Redman,C.W.;Wainscoat,https://www.wendangku.net/doc/6f9884025.html,ncet1997,350,485–487.

(31)Wright, C. F.;Burton,H.Hum.Reprod.Update2009, 15,139–151.

(32)Pathak,A.K.;Bhutani,M.;Kumar,S.;Mohan,A.;Guleria,R. Clin.Chem.2006,52,1833–1842.

(33)Fan,H.C.;Blumenfeld,Y.J.;Chitkara,U.;Hudgins,L.;Quake, S.R.Clin.Chem.2010,56,1279–1286.

(34)Tong,Y.K.;Jin,S.;Chiu,R.W.;Ding,C.;Chan,K.C.;Leung, T.Y.;Yu,L.;Lau,T.K.;Lo,Y.M.Clin.Chem.2010,56,90–98. (35)Fan,H.C.;Blumenfeld,Y.J.;Chitkara,U.;Hudgins,L.;Quake, S.R.Proc.Natl.Acad.Sci.U.S.A.2008,105,16266–16271.

(36)Hosono,N.;Kubo,M.;Tsuchiya,Y.;Sato,H.;Kitamoto,T.; Saito,S.;Ohnishi,Y.;Nakamura,Y.Hum.Mutat.2008,29,182–189.

(37)Fan,H.C.;Blumenfeld,Y.J.;El-Sayed,Y.Y.;Chueh,J.;Quake, S.R.Am.J.Obstet.Gynecol.2009,200(543),e541–547.